SOUTH AMERICAN PLANTS. 3 



and a pentangular border so characteristic of Physalis : in these 

 equivocal points of structure^ it appears to me we may call in the 

 aid of their general habit in order to determine the genus to 

 which they should be referred, for in Physalis the inflorescence 

 will be found to be universally 1 -flowered in each axil, while in 

 Saracha it is as uniformly more or less distinctly umbellate. 

 Thus likewise in AcnistuSy a genus with Cestrum-like flowers, we 

 have a very variable length of the tube of the corolla, which in 

 A. umbellatus is hardly distinguishable from the section CJm- 

 nesthes oi lochroma ; while in A, arborescens (the original Oestrum 

 cauliflorum of Jacquin, Hort. Schoenb. tab. 325)*the tube is so 

 short as to leave no possible distinction between this genus and 

 that called Witheringia by Kunth, as will be hereafter demon- 

 strated. 



Now, as will be hereafter shown, neither Witheringia so- 

 lanacea, nor the Columbian plant here alluded to as being so 

 closely allied to it, can be distinguished from Saracha ; they have 

 both a 5-partite calyx, a rotate corolla deeply cleft, stamens ari- 

 sing from triangular expansions originating at the base of its 

 short tube, and the fruit is a pisiform berry supported on a calyx 

 that does not materially increase in size ; the peduncle is bifur- 

 cate, and forms a 2-flowered umbel as in many species of Sara- 

 cha ; and to make this analogy still more complete, although the 

 stem is somewhat lignescent and perennial at base as in some 

 species of this last-mentioned genus, their branches are in like 

 manner herbaceous, and L^Heritier describes Witheringia sola- 

 nacea as possessing the same kind of large tuberose root as in 

 the Saracha jaltomata, Schlect. : for all which reasons I have no 

 hesitation in referring all these plants to one genus. 



Of the fruticose species hitherto included in Witheringia^ there 

 are evidently two distinct groups, the several Columbian species 

 enumerated by Kunth, and the Brazilian species of Martins : the 

 former are distinguished by having extra-axillary fascicles, gene- 

 rally of numerous, sometimes of very few flftwers, always upon 

 simple peduncles, and not umbellate as in Hebecladus ; the calyx 

 is always distinctly tubular, with an almost entire margin, and 

 five very minute distant teeth, not 5-partite as observed in Hebe- 

 cladus, Saracha, and Witheringia picta ; the corolla is tubular, 

 with a 5-partite border, not so decidedly long and infundibuli- 

 form as in Hebecladus and Acnistus ; the berry is small, seldom 

 exceeding the size of a peppercorn, and is supported on a small 

 persistent and nonaugescent calyx; it is not one-tenth the size of 

 the large oval berry inclosed within its increasing calyx, which is 

 seen in Witheringia picta ; the positive characters here alluded 

 to will be found to approach very closely to Acnistus, and to be 

 quite incompatible with the plants of the other group referred to. 



